Using random forest to test if two-wheeler experience affects driver behaviour when interacting with two-wheelers
Submission Date: 2019
Drivers are often at-fault in collisions with powered and unpowered Two-Wheelers (TW). In this paper, we propose a framework based on the random forest algorithm to investigate whether TW experience influences driver interactions with TWs. Sixty-nine drivers completed a 10-minute driving simulator session which included five interactions based on common car-TW crash types. The TWs were initially positioned in front of, or at right angles to, the driven vehicle. The proposed framework detected a statistically significant difference between drivers with TW experience and those without despite the small sample size.